The agent-native software factory
Guides on building, governing, and scaling an agent-native software factory, plus write-ups of the podcasts and talks behind the ideas.
Within Enterprise AI
Factory Private
Enterprise AI
Factory Private architecture inside your boundary
Factory Private architecture puts the control plane in your infrastructure. Plan the model, network, telemetry, and operating boundaries before a pilot.
Read articleAI Coding Agents
Enterprise AI
Fast inference changes the coding-agent workflow
Fast inference can shorten coding-agent feedback loops. Groq’s published Factory experience shows where model choice and end-to-end validation matter.
Read articleEnterprise AI
Security
Least-privilege access for coding agents
Least-privilege access for coding agents limits files, commands, networks, and credentials to what each task needs. Use this review checklist before granting autonomy.
Read articleAI Coding Agents
Engineering Management
Keeping long-running coding agents on track
Long-running coding agents need bounded milestones, externalized state, independent validation, and clear stop conditions. Use these controls for multi-day work.
Read articleFactory Private
Enterprise AI
Private AI for model research and post-training
Private AI keeps model research close to protected code, weights, and evaluation data. NVIDIA’s published Factory deployment shows why task trajectories matter.
Read articleFactory Private
Enterprise AI
Private AI package work needs durable context
Private AI package maintenance depends on repeatable builds and durable context. Chainguard’s published experience shows why long-running sessions deserve evaluation.
Read articleEnterprise AI
Factory Private
Enterprise AI needs approved engineering context
Enterprise AI depends on engineering context that security teams can approve. Nav’s published results connect reduced context switching with controlled access.
Read articleSoftware Factory
Enterprise AI
Self-improving software architecture in practice
Self-improving software architecture connects operational evidence to reviewed changes. Define what can change, where learning persists, and who controls the next release.
Read articleSoftware Factory
Enterprise AI
Self-improving software needs a failure record
Self-improving software needs evidence that a change removed a recurring failure. Factory’s published Signals research shows that loop with human approval before merge.
Read articleSoftware Factory
Enterprise AI
A software factory should learn from shipped work
A software factory connects incoming signals to reviewed changes and production feedback. Explore Factory’s existing architecture, automations, and validation resources.
Read articleFactory Private
Enterprise AI
On-prem AI development inside your security boundary
On-prem AI development needs more than a local agent. Place the control plane, model endpoint, build tools, and telemetry inside the approved boundary.
Read articleFactory Private
Enterprise AI
Self-hosted AI agents and control-plane ownership
Self-hosted AI agents require clear ownership of the control plane, session data, models, and updates. Compare the responsibilities before choosing a deployment.
Read article
